| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Research Articles: Therapeutics, Targets, and Development
Correction for chromosome-17 is critical for the determination of true Her-2/neu gene amplification status in breast cancer
1 Translational Research Unit and 2 Pathology Department, Bordet Institute, Brussels, Belgium; and 3 Human Genetics Department, BioMérieux, Marcy l'Etoile, France
Requests for reprints: Denis Larsimont, Department of Pathology, Jules Bordet Institute 1, rue Heger-Bordet, 1000 Brussels, Belgium. Phone: 32-2541-3115; Fax: 32-2541-3281. E-mail: denis.larsimont{at}bordet.be
Abstract
Purpose: Trastuzumab is the cornerstone for treatment of women with HER2-overexpressing breast cancer, both in the adjuvant and in the metastatic settings. The accurate assessment of HER2 is, therefore, critical to identifying patients who may benefit from trastuzumab-based therapy. This project aimed to determine the optimal scoring method for fluorescence in situ hybridization (FISH) assay. Methods: FISH assay was done on 893 samples of breast cancer. Three scoring methods were evaluated: Her2/CEP17
2, Her2>4, or Her2>6. Protein and gene expression were evaluated by immunohistochemistry (n = 584) and mRNA/assay/nucleic acid sequence–based amplification (NASBA; n = 90). Results: Samples were divided into five groups based on FISH results: disomic amplified and nonamplified, polysomic amplified, nonamplified, and discordant (10.8% of cases, mostly positive with Her2>4 scoring, but negative with the others). Her2/CEP172 and Her2>6 scoring methods showed the best association (a) with regard to FISH scoring (
= 0.906, P < 10–6) and (b) between FISH and immunohistochemistry (3+ as positive; > 0.650, P < 10–6) or NASBA ( > 0.536, P < 10–6). Polysomy had an effect on Her2 copy number (P < 10–6), but had no effect on protein and mRNA content. Therefore, within the discordant subgroup, for which additive Her-2 gene copies are due to high polysomy, protein and mRNA levels were similar to those of the nonamplified samples. For this subgroup, the best concordance between FISH/immunohistochemistry/NASBA was observed with the Her2/CEP17 ratio and Her-2>6 scoring (68% and 58% perfect matches, respectively). No perfect matches were observed using the Her2>4 scoring method. Conclusion: Correction for chromosome-17 is the method of choice for clinical practice; Her-2>6, but not Her-2>4, could be used as an alternative. [Mol Cancer Ther 2006;5(10):2572–9]
Introduction
The Her-2/neu oncogene, located on chromosome-17, encodes a transmembrane-tyrosine kinase receptor protein belonging to the epidermal growth factor receptor family. Protein overexpression, identified in 20% to 30% of human breast cancer tumors, is due to gene amplification in >90% of cases. Overexpression and amplification have been linked to poor prognosis and response to therapy with the anti–HER-2-humanized monoclonal antibody, trastuzumab (Herceptin), in patients with advanced metastatic breast cancer, when used either as single agent (1, 2) or in combination with chemotherapy (3). Recently, trastuzumab used in sequence or in combination with adjuvant chemotherapy, showed an impressive improvement in disease-free survival (4–7) and overall survival (5) in women with HER-2-overexpressing early breast cancer enrolled in five randomized clinical trials. Therefore, laboratory assessment of Her-2/neu status has become a basic procedure for the appropriate management of patients with breast cancer (8).
Determining the best method to identify patients who should be treated by trastuzumab-based therapy has been a source of controversy. Indeed, immunohistochemistry, which relies on several antibodies (monoclonal and polyclonal) and detection systems, produces contradictory results. HercepTest (Dako), a more standardized immunohistochemical method, was developed to overcome or decrease this variability, although with sometimes disappointing results (9). These concerns regarding the accuracy of immunohistochemistry using standard formalin-fixed paraffin-embedded tissue sections (10) have stimulated the use of fluorescence in situ hybridization (FISH) assay. Moreover, several studies have shown that FISH is a better predictor of prognosis and is a better determinant of trastuzumab eligibility compared with immunohistochemistry (2, 11, 12), as it excludes the majority of 2+ cases, which seem to have a low probability of response to trastuzumab treatment (1, 3). As a result, most laboratories either screen all cases by immunohistochemistry and triage-selected cases for FISH testing, or use FISH as the primary method for HER-2/neu testing. Therefore, the accurate assessment of Her-2/neu status based on gene amplification using FISH has become a critical factor in the management of patients with breast cancer.
Currently, there are two Food and Drug Administration–approved FISH assays with different definitions of Her-2/neu gene amplification: (a) the Ventana Inform test, which measures only Her-2/neu gene copies, and (b) the Vysis PathVysion test, which includes a chromosome-17 probe in a dual-color format. Tumors with an absolute Her-2 copy number of >4.0 per nucleus (Ventana), or with a ratio of 2.0 Her-2 gene copies to that of CEP17 (Vysis), are considered as amplified.
Several genomic situations can be observed (Fig. 1
). For disomic samples, the situation is usually easy; samples will be amplified or nonamplified, whatever the FISH scoring method used. The situation is more complex for polysomic samples (chromosome-17 copy number alteration) with >4 Her-2 gene copies because in some cases, a FISH status discrepancy is observed depending on the scoring method. On the one hand, considering only the absolute number of Her-2 gene copies to define amplification might overestimate the presence of gene amplification because 15% of cases with >4 Her-2 gene copies present a Her-2/CEP17<2 ratio because of polysomy (13); on the other hand, the proportion of samples identified as Her-2/neu-amplified by the Her-2/CEP17 ratio might be underestimated because a small proportion of cases (
3%) show increased numbers of Her-2 gene copies secondary to polysomy-17. It is conceivable that protein overexpression can result not only from an increased number of gene copies secondary to gene amplification, but also from concomitantly increased numbers of chromosome-17, with the consequence that patients whose tumors show these characteristics might also be potential candidates for trastuzumab treatment (13). Today, there is no definite consensus about the optimal scoring system for assessing Her-2 gene status. Some investigators (2, 11, 12, 14–16) have found correction for chromosome-17 to be critical for the determination of true gene amplification status as opposed to increased Her-2 gene copy number due to polysomy-17, whereas others (17, 18) believe that it is not required. Therefore, it is essential to establish a cutoff value or scoring system for Her-2 FISH amplification that takes into account this subset of patients with polysomy-17.
|
Materials and Methods
Patient Samples
The breast cancer samples used were sent to the Pathology Department of the Bordet Institute for routine clinical FISH assessment (13). For 893 paraffin-embedded samples, FISH assay was done using the PathVysion assay (Vysis; ref. 13). For 584 of 893 samples, HER-2 expression was also evaluated by immunohistochemistry (CB11, Novocastra; ref. 13). For the NASBA assay, which was done in collaboration with BioMérieux (Marcy l'Etoile, France), frozen samples containing >70% of invasive tumor cells were collected from 90 of 893 patients. This study was approved by the ethics committee of the Bordet Institute.
Criteria for FISH Positivity and Polysomy
Samples with a Her-2/CEP17 ratio of 2.0 or with an absolute Her-2 copy number per nucleus of >4.0 or >6.0 were considered as amplified. This last scoring system was used for the chromogenic in situ hybridization assay. As described previously (12), the cutoff points for chromosome-17 copy number alterations were estimated for each group according to published standards (14, 19–23) adapted for truncated nuclei on paraffin sections. Specimens with chromosome-17 copies in the range of 1.5 to 2.25 signals/cell were defined as having disomy-17. The other cases were considered to have aneusomy-17: either hypodisomy-17 (<1.5 signal/cell) or polysomy (>2.25/cell).
Duplex Real-time NASBA
Total RNA was extracted from five 10-µm frozen sections/patient case using the TRIZOL reagent according to the manufacturer's instructions (Invitrogen, Canada). RNA integrity and quantification were determined with the RNA 6000 Nano LabChip (Agilent, Germany).
Quantitative measurement of Her-2 mRNA was done with 50 ng of total RNA per duplex-NASBA reaction using cyclophilin B (PPIB) as the normalizing gene, as described previously (24). Forward and reverse primers for Her-2/neu and PPIB genes are: ERBB2-P1(T7) 5'aattctaatacgactcactatagggagaaggGAGCCAGCCCGAAGTtacgactcactatagggagaaggGAGCCAGCCCGAAGTCTGTA3', ERBB2-P2 5'TCTTAGACCATGTCCGGGAAA3', PPIB-P1(T7) 5'aattctaatacgactcactatagggagaaggCAGGCTGTCTTGACTGTCGTGA3', PPIB-P2 5'AGGAGAGAAAGGATTTGGCT3'. The 3' antisense primers are elongated with a T7-promotor recognition sequence indicated in lowercase characters. Molecular beacons are FAM-cgatcgGGAGGATGTG-CGGCTCGTACcgatcg-Dabsyl for the Her-2/neu gene and ROX-cgatcgGATCCAGGGCGGAGACTTCAcgatcg-Dabsyl for the PPIB gene. The stem sequences of the beacons are indicated in lowercase italic characters. GenBank accession numbers, MN_00448 for ERBB2 and M60857 for PPIB.
Statistical Analysis
The present work focused on determining the optimal cutoff to define Her-2/neu amplification and on its association with HER-2 expression at the transcriptional and translational levels. Moreover, the association between the various subgroups, defined on the basis of the FISH results and some clinicopathologic data were also evaluated. Standard descriptive analyses were conducted. The association between variables was assessed using the concordance coefficient (for binary variables), Mann-Whitney (for binary versus continuous variables) and 
2 tests. P < 0.05 was considered statistically significant. Statistical analyses were done using SPSS 11.5 (SPSS Inc., Chicago, IL).
Results
Polysomy-17 Is Responsible for Discordance in Her-2/neu Status as Defined by Different FISH Scoring Systems
The CEP17 and Her-2/neu gene copies, evaluated by FISH, were available for 893 samples (ref. 13; see Supplemental Data).4. For 96 of 893 cases (10.75%), Her-2/neu status was considered to be discordant according to the three FISH scoring methods, i.e. (a) the Her-2/CEP17 ratio, (b) the Her-2 copy >4 scoring, and (c) the Her-2 copy >6 scoring (chromogenic in situ hybridization criteria).
As reported in Table 1
, 78% (75 of 96) of the discordance was observed between Her-2>4 and the two other criteria, and 44% (42 of 96) between Her-2>6 scoring and the Her-2/CEP17 ratio. This explains why in the overall population, the best association between FISH scoring methods was observed between ratio Her-2/CEP17 and Her-2>6 scoring systems ( = 0.906, P < 10–6) compared with the association between Her-2>4 and the two other systems ( = 0.832 with ratio Her-2/CEP17 and 0.833 with Her-2>6; P < 10–6).
|
According to CEP17 and Her-2/neu copy number, samples were divided into five major groups: amplified polysomic, amplified disomic, nonamplified polysomic, nonamplified disomic groups, and discordant polysomic. The comparison of these groups revealed that polysomy-17 had an effect on Her-2 copy number, as gene copy number was statistically higher in polysomic versus disomic groups (P < 10–6). Additionally, a statistically higher CEP17 copy number was observed in polysomic amplified versus polysomic nonamplified samples (P = 0.005). In the discordant polysomic subgroup, which presented the highest CEP17 copy number (all P < 3 x 10–5), polysomy by itself accounted for a statistically higher Her-2/neu gene copy number when compared with the nonamplified polysomic subgroup (P < 10–6; Table 2 ).
|
No statistically significant difference could be observed between amplified polysomic and disomic subgroups or between nonamplified polysomic and disomic subgroups in terms of HER-2 protein expression (0, 1+, 2+ versus 3+; Table 2). Importantly, there was no statistically significant difference observed between polysomic nonamplified and discordant samples (P = 0.21). Therefore, additional Her-2 gene copies induced by polysomy mainly in the discordant cases (compared with the nonamplified polysomic subgroup; P < 10–6) were not associated with HER-2 overexpression. Nevertheless, in 3 of 161 IHC3+ samples, HER-2 overexpression (as confirmed by Herceptest) was due to polysomy-17. It is noteworthy that for these cases, the CEP17 copy number exceeded 10 copies per cell, which resulted in more than 19 Her-2 gene copies/cell as confirmed by the chromogenic in situ hybridization assay (data not shown).
When considering HER-2 IHC3+ as positive, the association between immunohistochemistry and FISH data dichotomized according to the three scoring systems was 0.689 using Her-2>6, 0.650 using the Her-2/CEP17 ratio and 0.556 using Her-2>4 (all P < 10–6). Concordances between immunohistochemistry and FISH data, according to the three scoring systems, are summarized in Table 3
for the discordant polysomic subgroup. With Her-2>4 FISH scoring, 42 samples classified as amplified were not scored as IHC3+. In contrast, with the two other scoring systems, only one to four samples classified as nonamplified were scored as IHC3+. Therefore, concordance between FISH and immunohistochemistry was observed in 10.6% of samples using Her-2>4 scoring, in 87% of samples using Her-2>6 scoring, and in 81% of samples using the Her-2/CEP17 ratio.
|
association between immunohistochemistry and NASBA assay was 0.465 (P < 10–4). In terms of Her-2 mRNA content, we observed no statistically significant difference between amplified polysomic and disomic groups, whereas there was a statistically significant difference between amplified groups and the nonamplified polysomic subgroup (both P < 4 x 10–5). Interestingly, we found no statistical difference between polysomic nonamplified and discordant cases. Therefore, additional Her-2 gene copies induced by polysomy in the discordant cases were not associated with higher gene expression.
The Her-2/CEP17 ratio scoring (Z = –5.376) and the Her-2>6 definition (Z = –5.392), when compared with the Her-2>4 (Z = –3.992) criterion (all P < 7 x 10–5), provided the best association between NASBA assay and FISH. Using a cutoff of >0.078 (ratio Her-2/PPIB) for NASBA, the associations between FISH and NASBA were 0.564 for Her-2>6, 0.536 for ratio Her-2/CEP17, and 0.325 for Her-2>4 (all P < 3 x 10–5). Concordance between the three FISH scoring methods and NASBA for the subgroup of discordant breast cancer is summarized in Table 3. With the Her-2>4 FISH scoring, 19 patients classified as amplified by FISH did not overexpress mRNA; this number decreased to three and one patient(s) with the two other scoring systems. In contrast, three and four patients with high mRNA levels (all IHC2+) were classified as nonamplified by FISH using the Her-2>6 and the Her-2/CEP17 ratio criteria. Therefore, 17%, 74%, and 83% of samples were concordant according to NASBA and the Her-2>4, Her-2>6, and Her-2/CEP17 ratio scoring systems, respectively.
Perfect concordance between FISH, immunohistochemistry, and NASBA was observed in 68% of cases with the Her-2/CEP17 ratio criteria, in 58% with the Her-2>6 criteria, and in 0% with the Her-2>4 criteria. Therefore, correction for chromosome-17 seemed to be the optimal scoring system for FISH amplification evaluation. Her-2>6 scoring could be used as an alternative criterion, even if among the four of five samples with conflicting status according to Her-2>6 or ratio Her-2/CEP17 criteria, no protein (IHC<3+) and no mRNA expression (except one) was detected. The Her-2>4 criterion seemed to be the worst scoring system (Fig. 2 ). We observed that the minimal cutoff for absolute Her-2 copy number must be more than six copies in order to obtain a good association between FISH and NASBA or immunohistochemistry, the association with immunohistochemistry or NASBA being increased by 1.5 with a cutoff of 6 rather than 4.
|
|
Discussion
To determine the optimal FISH scoring system to select patients that might benefit from a trastuzumab-based treatment, three criteria for FISH positivity were considered, and the ensuing status was then compared with immunohistochemistry and mRNA/NASBA data, two techniques evaluating HER-2 expression status at a different level (translation and transcription, respectively) and whose results seemed to be significantly associated. Similar to previous studies (26, 27), we found a good concordance between Her-2/CEP172 ratio and Her-2>4 FISH scoring methods, for disomic and low polysomic cases, as well as for cases with increased Her-2 gene copies consecutive to gene amplification. Problematic cases, discrepant for FISH scoring result, are samples with no gene amplification but increased Her-2 gene copies secondary to high polysomy-17 (>3 copies per cell) in agreement with Lal et al. (27). Here, these samples corresponded to 10% of the patients for which trastuzumab would be given based on Her-2>4 criterion, but not on the Her-2/CEP17 ratio criterion.
We clearly observed that polysomy-17 had an effect on Her-2 copy number. On the one hand, a statistically higher Her-2 gene copy number was found in polysomic groups, but on the other hand, the polysomic amplified group presented a higher number of CEP17 copy number compared with the polysomic nonamplified group, as previously reported (20, 28). The challenge was to determine whether additional Her-2 gene copies consecutive to polysomy-17 might be sufficient to induce HER-2 protein overexpression, the target of trastuzumab, with the consequence being that these patients might be candidates for trastuzumab treatment. As previously reported (15, 20, 29, 30), we clearly showed that, in the absence of Her-2 gene amplification, polysomy-17 resulted in a modest increase of Her-2/neu gene copies and, consequently, a rare increase of HER-2 protein and mRNA content, except for a very small number of cases with more than 10 copies of CEP17 per cell. As a consequence, correction for chromosome-17 seems to be critical for the determination of true Her-2 gene status in breast cancer. These results were corroborated by the fact that the best association between FISH and immunohistochemistry, as reported by other studies (27, 29), as well as between FISH and NASBA, seemed to be with Her-2/CEP17 ratio criteria. Wang et al. (29) reported that FISH assay using the Her-2/CEP17 criterion achieved higher concordance with image analyzer–assisted immunohistochemical quantification than did the Her-2>4 criterion. Lal et al. (27) observed that more immunohistochemistry-negative tumors were FISH-positive with the Her-2>4 criterion. Therefore, the Her-2/CEP17 ratio, which considered most of these discordant cases as nonamplified, seemed to be the better scoring system to reflect the real biology of the tumor.
As an alternative to the Her-2/CEP17 ratio, absolute Her-2 gene copy >6 criterion could be used as a good scoring system. A recent report recommended changing the cutoff mean Her-2 signal number to 6 for the Ventana-Inform assay to account for the discrepant cases due to polysomy-17 (31). We effectively confirmed that the minimum number of Her-2 gene copies necessary to obtain a good correlation between FISH, immunohistochemistry, and NASBA must be >6. This scoring system might also be useful mainly for cases with more than 10 CEP17 copies per cell and IHC3+. This hypothesis must be confirmed on a higher number of samples.
The comparison of the biological characteristics showed that discordant polysomic samples were biologically similar to polysomic nonamplified samples. Indeed, like others (28, 32), we observed that Her-2 amplification was associated with a lower proportion of estrogen receptors, as opposed to the nonamplified and discordant samples. We also reported that tumoral genomic alterations, either amplification or chromosome-17 polysomy, were associated with proliferation, although, no correlation was observed with the histologic grade, probably due to the high number of cases with unknown status. Our results were in agreement with a recent report from Fridlyand et al. (33), which also showed a correlation between chromosomal level instability as assessed by comparative genomic hybridization and grading. They discriminated three breast tumor subtypes according to their genomewide DNA copy number profiles. At the low end of chromosomal level instability were ER-positive tumors, which were exclusively of moderate or well-differentiated grade and whose patients did not recur. At the other extreme of genome instability were tumors with many low-level copy number aberrations, which were ER negative, of high grade, and associated with significantly worse outcome compared with the other groups. The third group comprised both ER-positive and -negative tumors and was characterized by the presence of low level gains and losses and recurrent amplifications (33).
Recent DNA microarray profiling studies on breast tumors have identified distinct subtypes of breast carcinoma associated with different clinical outcomes (34–37). On the basis of these studies, immunohistochemical classification of breast carcinomas has been proposed (38), wherein Her-2 status evaluation is critical for all subgroup assignations. Consequently, apart from the identification of patients that might benefit from trastuzumab-based treatment, the definition of FISH scoring system is important for the classification of breast carcinomas in a subgroup of poor or good clinical outcome. Among the discordant samples according to FISH criteria, samples identified as HER-2-expressing with Her-2>4 scoring would be classified as basal-like or luminal-like depending on the ER status and according to the two other criteria. Therefore, a sample with a potentially good clinical prognosis (luminal-like) could be classified in a subgroup of poor prognosis with all the therapeutic consequences that it may imply.
Three recently published randomized adjuvant trastuzumab trials have considered both immunohistochemistry and/or FISH to assess HER-2 protein overexpression and/or gene amplification. As the effectiveness of a targeted therapy depends on the identification of potentially responsive patients and to avoid lack of agreement among studies, patients should be registered based on HER-2 expression evaluation in reference laboratories or by FISH. Nevertheless, even with the use of FISH, response rates may differ according to the FISH criteria used. The definition of a unique FISH scoring system is therefore crucial for the comparison of response rate among studies, and correlation between the scoring system and treatment efficacy must be assessed in trastuzumab-treated patients in order to determine which scoring system accounts for the best response rate. Moreover, with the advent of the significantly positive first results of these trials, it has become crucial to limit the number of false-positive cases, as trastuzumab is associated with the risk of severe cardiotoxicity. By limiting this number, the use of the Her-2/CEP17 ratio would effectively identify potentially responsive patients and limit the number of patients exposed to potential toxic effects without any treatment benefit.
To conclude, we showed that the correction for chromosome-17 is the optimal scoring system for FISH amplification evaluation. Alternatively, an absolute Her-2 gene copy number of >6 could be used. Our results suggest that the CEP17 polysomic subset of carcinoma without Her-2/neu gene amplification behaves similarly to nonamplified tumors. Nevertheless, this hypothesis still needs to be validated in trastuzumab-treated patients.
Acknowledgments
We thank "Les Amis de L'Institut Bordet" and "The Fond Lambeau-Marteaux" for research support.
Footnotes
Grant support: C. Desmedt was supported by a grant from the Fonds National de la Recherche Scientifique.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Note: L. Dal Lago, V. Durbecq, and C. Desmedt should be considered as co-first authors.
4 Supplementary materials for this article is available at Molecular CancerTherapeutics Online (http://mct.aacrjournals.org). 
Received 3/ 8/06; revised 6/ 8/06; accepted 8/16/06.
References
T) with doxorubicin and cyclophosphamide followed by docetaxel and trastuzumab (ACTH) with docetaxel, carboplatin and trastuzumab (TCH) in HER2 positive early breast cancer patients: BCIRG 006 study [abstract 1]. Breast Cancer Res Treat 2005;94. Shak S, Herceptin Multinational Investigator Study Group. Overview of the trastuzumab (Herceptin) anti-HER2 monoclonal antibody clinical program in HER2-overexpressing metastatic breast cancer. Semin Oncol 1999;26:71–7.[Medline] Jacobs TW, Gown AM, Yaziji H, et al. Specificity of HercepTest in determining HER-2/neu status of breast cancers using the United States Food and Drug Administration–approved scoring system. J Clin Oncol 1999;17:1983–7.This article has been cited by other articles:
|
|
 |
|
  L. Arnould, P. Arveux, J. Couturier, M. Gelly-Marty, C. Loustalot, F. Ettore, C. Sagan, M. Antoine, F. Penault-Llorca, B. Vasseur, et al. Pathologic Complete Response to Trastuzumab-Based Neoadjuvant Therapy Is Related to the Level of HER-2 Amplification Clin. Cancer Res., November 1, 2007; 13(21): 6404 - 6409. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. C. Wolff, M. E. H. Hammond, J. N. Schwartz, K. L. Hagerty, D. C. Allred, R. J. Cote, M. Dowsett, P. L. Fitzgibbons, W. M. Hanna, A. Langer, et al. American Society of Clinical Oncology/College of American Pathologists Guideline Recommendations for Human Epidermal Growth Factor Receptor 2 Testing in Breast Cancer J. Clin. Oncol., January 1, 2007; 25(1): 118 - 145. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
